RELIABLE DISTRIBUTED FAULT DIAGNOSIS USING REDUNDANT DIAGNOSERS

The paper deals with distributed fault diagnosis of partially observed discrete event systems (DES) modelled by Interpreted Petri Nets. The diagnosability property is studied for the set of resulting sub-models of the distributed Interpreted Petri Net. An architecture for reliable distributed on-line diagnosis is presented; it includes a set of redundant diagnosers that handle the corresponding reduced sub-model and other sub-models corresponding to the neighbour diagnosers.